THE EFFECTS OF TAXATION ON THE LOCATION · PDF fileNational Tax Journal, September 2011, 64 (3), 817–838 THE EFFECTS OF TAXATION ON THE LOCATION DECISION OF MULTINATIONAL FIRMS:

National Tax Journal, September 2011, 64 (3), 817–838

THE EFFECTS OF TAXATION ON THE LOCATION DECISION OF MULTINATIONAL FIRMS:

M&A VERSUS GREENFIELD INVESTMENTS

Shafi k Hebous, Martin Ruf, and Alfons J. Weichenrieder

In this study, we estimate the impact of differences in international tax rates on the probability of choosing a location for an affi liate of a multinational fi rm. In particular, we distinguish between the tax sensitivity of Greenfi eld and Mergers and Acquisitions (M&A) investments. Based on a novel fi rm-level dataset on German outbound foreign direct investment (FDI), we fi nd evidence that location decisions of M&A investments are less sensitive to differences in tax rates than location decisions of Greenfi eld investments. According to our logit estimates, after controlling for fi rm and country-specifi c characteristics, the tax elasticity for Greenfi eld investments is negative and in absolute value signifi cantly larger than that associated with M&A investments. This fi nding is consistent with (partial) capitalization of taxes in the acquisition price when the FDI project takes the form of a M&A.

Keywords: FDI, international taxation, location, Mergers and Acquisitions, Greenfi eld investment

JEL Codes: H25, H73, F23

I. INTRODUCTION

The role of taxation in explaining the international allocation of investment has been a subject of immense theoretical and empirical scrutiny. Undoubtedly, several fac-

tors affect the location decision of a multinational fi rm. The tax system of the potential host economy is one of these factors, positioning this issue at the intersection of sev-eral branches of economics: public fi nance, international economics, and international business.

Shafi k Hebous: Faculty of Economics and Business Administration, Goethe University, Frankfurt am Main, Germany ([email protected])Martin Ruf: Business School, Mannheim University, Mannheim, Germany ([email protected]) Alfons J. Weichenrieder: Faculty of Economics and Business Administration, Goethe University, Frankfurt am Main, Germany, and Vienna University of Economics and Business, Vienna, Austria, and CESifo, Munich, Germany ([email protected])

National Tax Journal818

Earlier empirical studies based on aggregate fi gures on foreign direct investment (FDI) suggest that high taxes negatively impact the fl ow of FDI (de Mooij and Ederveen, 2003; Hines, 1999). Recent empirical work exploits data at the fi rm level to estimate the effects of taxes on the decision of where to locate an affi liate of a multinational fi rm. By relying on microeconomic data, these studies provide information on the impact of taxes on the behavioral responses of multinational fi rms to international differences in taxation while adequately accounting for differences across fi rms and industries. How-ever, most existing empirical studies treat different FDI modes as homogenous projects, thus missing one crucial piece of information — the distinction between Mergers and Acquisitions (M&A) and Greenfi eld investment projects (new ventures).1

In this study, we account for the mode of investment in estimating the effect of differences in international tax rates on the probability of choosing a location for an affi liate of a multinational fi rm. Specifi cally, there are reasons to expect that the location decision of Greenfi eld investments is more sensitive to differences in international tax rates than the location decision of M&A projects. First, if the potential M&A project is located in a high tax country, a portion of the tax burden might be capitalized, reduc-ing the acquisition price. This capitalization effect is less pronounced in the case of a Greenfi eld investment and suggests that M&A investments should react less to high taxes than Greenfi eld investments. Second, M&A decisions depend on the availability of appropriate targets. In principle, the set of potential locations for establishing a new plant might be larger than the set of locations of potential target fi rms that might be acquired. This may make the multinational fi rm less constrained in optimising over the location decision of the new venture.

In our econometric analysis, we employ detailed fi rm-level data on German outbound FDI covering about 3,600 fi rms from 2005–2007. The most valuable feature of this dataset for our purposes is that since 2005 the German investor has to report whether a new FDI project is a Greenfi eld or M&A project. This enables us to directly identify the mode of investment at entry. In our sample, 34 percent of fi rms enter the host economy as a Greenfi eld project. The United States is the largest receiver of new FDI entries, with a share of about 11 percent of total new German outbound FDI projects.

Our main fi ndings are summarised as follows. First, if we do not distinguish between the modes of entry, high tax rates reduce the probability that a location will be chosen by a German FDI investor for its new affi liate. This fi nding is consistent with the results of Devereux and Griffi th (1998) on the location of U.S. multinationals abroad. Our second fi nding however reveals that Greenfi eld investments are signifi cantly more elastic to international taxation than M&A investments. According to our logit estimates, after controlling for fi rm and country-specifi c characteristics, an increase in the statutory corporate income tax rate of 10 percent reduces the probability of choosing a coun-try to host a Greenfi eld investment by about 6.4 percent. The tax elasticity for M&A investments however, although negative, is signifi cantly smaller and roughly half the tax elasticity of Greenfi eld investments.

1 See Devereux (2007) for a survey.

M&A Versus Greenfi eld Investments 819

Our study is the fi rst to document this empirical fi nding using outbound microeco-nomic data. Swenson (2001) examines the composition of FDI within the United States and fi nds that Greenfi eld investments are more deterred in high tax states than are M&A.2 The idea that the impact of taxation may depend on the modes of FDI dates back to Auerbach and Hassett (1993) who argue that tax reforms can alter the incentives for investing in the acquisition of old capital rather than investing in new capital. Becker and Fuest (2008) present a theoretical model of tax competition in which an increase in the tax rate increases the number of M&A investments and decreases the number of Greenfi eld investments. Huizinga and Voget (2009) examine the impact of double taxation on organizational structure following cross-border M&A activities. They fi nd that countries with high levels of international double taxation are less likely to host the new parent fi rm after a merger or acquisition has occurred.3

Recent contributions in the literature on the theory of international trade predict that fi rm- and project-specifi c characteristics play a major role in determining the mode of entry. Nocke and Yeaple (2008) and Raff, Ryan, and Stähler (2009) show that more productive fi rms tend to enter a foreign market as a Greenfi eld rather than a M&A investment. Consistent with this result, Andersson and Svensson (1994) fi nd that fi rms with high technological skills and research and development intensity are more likely to make Greenfi eld investments. Neary (2007) introduces an oligopolistic market structure in a general equilibrium framework and shows that fi rms acquire their high cost rivals. In Nocke and Yeaple (2007), fi rms undertake M&A to complement their strengths.

The differential impact of high tax rates on the composition of FDI is also of economic interest as Greenfi eld investments and M&A activities can have different implications for the host economy. M&A activities change the pattern of ownership rights but may have small effects on the international allocation of capital, the levels of production capacity, and labor demands. Furthermore, the pattern of productivity spillovers from multinational affi liates to domestic fi rms through knowledge spillovers and changes in the competitive structure of the host economy may differ between M&A and Greenfi eld projects (Balsvik and Haller, 2007).

This study proceeds as follows. In Section II, we present our empirical approach and describe the German fi rm-level FDI dataset. We report the main results in Section III, and present a robustness analysis in Section IV. Finally, we conclude in Section V.

II. EMPIRICAL METHODOLOGY AND DATA

A. Empirical Approach

Our econometric analysis is cross-sectional focusing on parent fi rms that decide to establish a new affi liate in a host country, and is based on the rich literature on FDI

2 Swenson (2001) fi nds that high taxes have a positive but insignifi cant effect on M&As. 3 The infl uence of taxes on the ownership of foreign affi liates has recently been emphasized by Desai and

Hines (2003) and Becker and Fuest (2010).


location choice. The parent fi rm selects the location associated with the highest expected profi t.4 The profi t functions (∏i ) associated with each location i = 1, 2, …, n are

(1) i i i i ij

iR C I= ⎡⎣⎡⎡ ⎤⎦⎤⎤( )i− ( )i ,i i ijR Ci i I⎡⎡ ⎤⎤ −)i (

where j denotes either a Greenfi eld or a M&A investment. The variable τi is the corpo-rate tax rate in location i. Revenue is Ri and the cost function is Ci, which can depend on several factors such as output, cost of labor, agglomeration and other external economies, etc. The term Ii

j(τi) captures the initial cost of the FDI in the case of Green-fi eld projects and the purchase price in the case of M&A projects. The optimal levels of output can be derived by solving the system of fi rst order conditions. Optimal profi ts (∏i

* ) can be computed by substituting the optimal levels of output in the corresponding profi t functions. Ultimately, however, our focus here is on the effects of taxation on location decisions of M&A versus Greenfi eld investments.

High taxes in the host economy lower future cash fl ows from the cross-border invest-ment in both types of projects and thus deter both types of FDI. However, in the case of a M&A project the acquisition price is a function of the tax rate. The value of a fi rm is the present value of the cash payoffs received by the claim holders of the fi rm (Healy and Palepu, 2007). Corporate taxation reduces the value of a fi rm (Modigliani and Miller, 1963). Consequently, a high tax rate may reduce the fi nal price paid by the buyer for a potential acquisition. Huizinga, Voget, and Wagner (2008) provide empirical evidence on the existence of a tax capitalization effect in takeover premiums. Hence, if part of the tax is capitalized in the purchase price, then ∂Ii

M&A(τi)/∂τi < 0. This capitalization effect, however, is expected to be smaller in the case of a Greenfi eld investment since many mobile capital goods can be purchased at world prices. Thus, the capitalization of taxes in the acquisition price suggests that the impact of taxes on the location decision is mitigated in the case of M&A investments as compared to the case of Greenfi eld investments.

Empirically, we observe the binary latent variable

(2) yi n l i

k lk l k i

,,

*

,

* , , ,... ,=

≠1i* *1 ik i , 2

0

if and

otherwi

k l* >

sess ,

⎧⎨⎪⎧⎧⎨⎨⎩⎪⎨⎨⎩⎩

where the subscript k denotes fi rms. The profi tability of location l depends, among other things, on the statutory corporate income tax rate in location l (τl)

(3) Πk l l k l k,lkk k lA,

*

,( & ) ,Φ k lk k lΦα βτ l τ l k lkΦ kΦ lk

where α is the intercept, β and γ are the coeffi cients of interest, Φ is the vector of coeffi cients corresponding to the control variables, and εk,l is a residual. To investigate

4 Since we observe a FDI decision, we focus on the decision where to invest rather than the decision whether to stay home or go abroad; see Markusen (2002) for a general equilibrium treatment of these decisions.


whether or not the two modes of investment react differently to taxes, we include the interaction term M & Ak × τl, where M & A is a dummy variable that takes the value 1 if the investment takes the form of a M&A and 0 if the investment is a Greenfi eld project. The multiplication of the M & A dummy and the tax rate is required to compute interaction effects.5 The vector xk,l is a vector of alternative-specifi c control variables that includes country-specifi c controls, i.e., controls that vary across countries but not fi rms, as well as fi rm-country controls that vary across countries but are fi rm specifi c. In some specifi cations, we also take into account industry-specifi c effects.

The probability of fi rm k choosing location l is given by

(4) p e

ek l

i

n

k l

k i

,

,*

,*

.=

=∑

Π

Π

1

This logit model is estimated using the maximum likelihood method. The tax vari-able τl is central in our analysis and its impact is expected to be negative and larger in absolute value for Greenfi eld than M&A investments. The identifi cation is based on cross-sectional variation in tax rates. Data on tax rates are taken from Mintz and Weichenrieder (2010). Table 1 displays the statutory corporate income tax rates for the countries in our sample. In 2006, for example, tax rates range from a minimum of 12.5 percent in Ireland to a maximum of 40.7 percent in Japan. The choice of the statutory tax rate might be particularly appropriate for M&A, if income shifting possibilities help motivate the transactions. Additionally, our focus on the statutory corporate income tax rates is further motivated by the empirical evidence presented in Büttner and Ruf (2007), who fi nd that the effective average tax rate has considerably less explanatory power than the statutory rate. Nonetheless, in the robustness analysis, we also examine the effects of the effective average corporate income tax rate on the location decision.

The vector xk,l includes several variables that are related to the probability locating multinational fi rm will locate an affi liate in a host economy. First, the previous pres-ence of a fi rm in location l is captured by the number of affi liates already operating in the host economy (no. affi liates). The presence of a fi rm in an economy may increase the probability of selecting that economy again. Second, the total fi xed and intangible assets invested by the parent fi rm in location l (total assets) accounts for the size of the prior investment of the parent fi rm in location l. Third the level of development of the host economy is captured by its GDP per capita (gdp capita). Fourth, the market size of the host economy is captured by its population (population). The level of development and market size both are expected to have positive effects on the probability of enter-

5 However, the dummy variable M & A without interaction is not included in the regression because it does not vary across the alternatives available for a fi rm. If, for example, the German parent acquires a fi rm in a country, then the dummy M & A takes the value 1 and the variable y also takes the value 1 for this country-fi rm observation. However, the M & A dummy also takes the value 1 for all other country-year observations corresponding to this location choice (for which the variable y takes the value 0).


ing. Fifth, local labor market conditions are captured by the labor freedom component of the Heritage index of economic freedom (labor freedom). This proxy is computed based on four factors: minimum wages, rigidity of hours, diffi culty of fi ring redun-dant employees, and cost of fi ring redundant employees.6 The labor freedom index is expected to be positively related to the probability of choosing a location l; the higher

Table 1Statutory Corporate Income Tax Rates

(Percent)

Country 2005 2006 2007 Country 2005 2006 2007Argentina 35 35 35 Malaysia 28 28 27Australia 30 30 30 Malta 35 35 35Austria 25 25 25 Mexico 30 29 28Belgium 34 34 34 Morocco 35 35 35Brazil 34 34 34 Netherlands 31.5 29.6 25.5Bulgaria 15 15 10 New Zealand 33 33 33Canada 34.4 34.2 34.1 Nigeria 30 30 30Chile 17 17 17 Norway 28 28 28China 33 33 33 Philippines 32 35 35Colombia 35 35 34 Poland 19 19 19Croatia 20.3 20.3 20 Portugal 27.5 27.5 25Czech 26 24 24 Romania 16 16 16Denmark 28 28 28 Russia 24 24 24Egypt 20 20 20 Saudi Arabia 20 20Finland 26 26 26 Singapore 20 20 20France 33.8 33.3 33.3 Slovakia 19 19 19Greece 32 29 25 Slovenia 25 25 23Hong Kong 17.5 17.5 17.5 South Africa 37.8 36.9 36.9Hungary 16 16 16 Spain 35 35 32.5India 36.6 33.7 34 Sweden 28 28 28Indonesia 30 30 30 Switzerland 21.3 21.3 21.3Ireland 12.5 12.5 12.5 Taiwan 25 25 25Italy 37.3 37.3 37.3 Thailand 30 30 30Japan 40.7 40.7 40.7 Turkey 30 30 20Kenya 30 30 30 UK 30 30 30Korea 27.5 27.5 27.4 Ukraine 25 25 25Lithuania 15 15 15 United States 39 39 38.6Luxembourg 30.4 29.6 29.6 Venezuela 34 34 34Source: Mintz and Weichenrieder (2010).

6 See Miller and Holmes (2009) for detailed information on the Heritage index.


the fl exibility of the labor market, the higher the probability of entering the economy.7 Sixth, the distance between Germany and the location of the FDI project is captured by (distance). The inclusion of this variable is in the tradition of the gravity literature. It captures transport (trade) costs and may also capture investors’ information on mar-ket conditions in the host economy. Distance is typically associated with a negative estimated coeffi cient. Seventh, the openness of the host economy to international trade is captured by the ratio (importsl + exportsl)/gdpl (openness). This proxy may capture two opposing factors. For example, the tariff-jumping argument suggests that the prob-ability of entering a relatively closed economy is rather high as the multinational fi rm attempts to get access to the market. At the same time, economies that are more open to international trade may be more open to international investments. Finally, the quality of a country’s institutions is captured by a corruption index (corruption). We employ the Corruption Perception Index of Transparency International. We redefi ne it such that a high value of this index indicates a high level of corruption.

Further, we include industry dummies in some specifi cations. These dummies account for difference across industries and potential industry-specifi c economies of scales. All level variables are expressed in terms of natural logarithms. In the robustness analysis, we examine the effects of the effective tax rate and further country-specifi c character-istics, such as research and development (R&D) intensity and the ratio of the market capitalization of listed fi rms to GDP. The data appendix provides a detailed description of the variables used in the study.

B. Firm Data and Descriptive Statistics

German foreign trade and payments regulations oblige all German fi rms and individu-als investing abroad who meet the reporting requirements to report key information such as balance sheet items as well as the economic sectors of the parent fi rm and its affi liates. In contrast to several other fi rm-level data sources, a valuable feature of this dataset is the inclusion of the entire population of FDI fi rms, rather than only listed or “big” fi rms. Since 2005 German investors are required to report whether a new investment is a Greenfi eld or M &A project.8 This is a novel piece of information that enables us to directly identify the mode of entry in our empirical investigation. We exclude from our sample banks, fi nancial institutions, and non-profi t institutions since they may face special tax rules.

The data cover the years 2005, 2006, and 2007. Figure 1 displays the number of new entries in each year. In total, 2,321 new cross-border M&A projects and 1,306

7 Some studies incorporate the average wage as a proxy for the labor market situation. However, in most studies this variable turned out to be insignifi cant, as in Devereux and Griffi th (1998) and Head and Mayer (2004). Furthermore, unfortunately, data on average wages are not available for many countries in our sample.

8 The investor has to characterize outbound FDI in one of four ways: (1) new entry Greenfi eld project, (2) new entry M&A project, (3) already existing fi rm that was reported in the previous year, or (4) fi rst time satisfying the reporting requirements (the fi rm existed last year but was not reported). The fi rst two options are the new entrants. Further details on the reporting requirements and German FDI dataset can be found in Lipponer (2008).


Greenfi eld investments are reported. In 2005, Greenfi eld investments constitute about 35 percent of total new entries. A similar pattern occurs in 2006 and 2007, with shares of 36 percent and 38 percent, respectively.

Table 2 shows that 63 percent of new entries take place in Europe, while the United States receives the largest single share of new entries (10.6 percent). Although 67.1 percent of entries in the United States are M&A investments, the United States has the second largest share of worldwide Greenfi eld entries (11.1 percent). China is the largest recipient of Greenfi eld investments with a share of about 15.9 percent of total German Greenfi eld entries worldwide. Table 2 also shows the regional pattern of FDI destina-tions. Several countries that share a common border with Germany such as Austria, Switzerland, the Netherlands, and Poland are among the top 10 recipients of new FDI projects. This pattern, consistent with the widely-recognized border effect, leads us to include a variable measuring the distance between Germany and the host country.

In terms of the size of the new investment, Figure 2 shows that the average fi xed and intangible assets of a FDI project vary across the modes of entry and locations. For example, in 2007 the average asset ratio of a M&A project allocated outside of Europe amounts to over 60 million euro, whereas this average for a M&A project within Europe is about 23 million euro.

Table 3 provides the means, standard deviations, 5th percentiles, and 95th percentiles of the variables that are used in our study, for both types of FDI. The fi gures are similar across both modes of investments. It should be noted that the statutory tax rate is highly

0

100

200

300

400

500

600

700

800

900

2005 2006 2007

M&A Greenfield

Figure 1Number of New Investments

Table 2German Outbound New FDI Entries, 2005–2007

(Percent)

Share in M&A Share M&A Share Greenfi eld ShareTotal New in New in Worldwide in Worldwide

Entry Entrants M&A Entrants Greenfi eld EntrantsEurope 63.0 73.0 66.9 54.4Outside Europe 37.0 61.4 33.1 45.6

Austria 5.4 71.0 5.6 5.0Belgium 2.5 84.0 3.1 1.3China 7.9 36.8 4.3 16.0France 6.4 83.3 7.8 3.4Italy 3.6 83.0 4.3 1.9Poland 3.5 59.2 3.0 4.5Russia 2.3 38.2 1.3 4.5Switzerland 4.2 77.6 4.7 3.0The Netherlands 4.9 76.6 5.4 3.6UK 8.7 86.1 10.9 3.9United States 10.6 67.1 10.4 11.1Notes: The reported fi gures are shares in the total number of entries. The table includes only the 10 most important host economies in our sample.

0

10

20

30

40

50

60

70

2005 2006 2007

Mill

ion

non-Europe Greenfield non-Europe M&A Europe Greenfield Europe M&A

Figure 2Average Fixed and Intangible Assets


positively correlated with the effective average tax rate (0.94). The correlation between the corruption level and the statutory tax rate is very low.

III. EMPIRICAL RESULTS

Table 4 presents our main results. The signs of the estimated coeffi cients on country-specifi c variables provide the directions of the effects of these variables on the probability a multinational fi rm will choose a location l for its investment. However, coeffi cients in non-linear models are not equivalent to marginal effects. Therefore, we report in Table 4 the corresponding estimated elasticities. The underlying logit estimation results for computing theses elasticities are not reported. In columns 1 to 3 we constrain the effects of taxation to be the same across modes of FDI. The difference between columns 1 and 2 is the inclusion of year dummies in column 2. To capture potential industry-specifi c location preferences, we re-estimate the benchmark model but include industry-specifi c dummies. The results are reported in column 3. As expected, the estimated tax rate elasticity is negative and signifi cant in all specifi cations. This fi nding is consistent with studies that perform similar exercises examining the location decisions of foreign affi li-ates (Büttner and Ruf, 2007; Egger et al., 2009). Also, this fi nding is consistent with recent results obtained from aggregate FDI fl ow fi gures, such as Djankov et al. (2010).

However, such specifi cations ignore the heterogeneous modes of investment. In the remaining specifi cations in Table 4, we allow taxation to have different impacts depend-

Table 3Descriptive Statistics

MeanStandard Deviation 5th Percentile 95th Percentile

Statutory tax rate 0.27 0.06 0.16 0.37

Effective tax rate 0.27 0.08 0.16 0.42

No. affi liates 0.16 0.42 0 1.09

Total assets 1.31 3.21 0 9.59

GDP capita 9.85 0.76 8.31 10.76

Population 16.91 1.62 14.51 19.51

Openness 1.05 0.82 0.3 3.02

Distance 7.86 1.19 6.14 9.41

Labor freedom 0.65 0.16 0.4 0.95

Corruption –5.77 2.37 –2.40 –9.39

Market capitalization 0.96 0.81 0.27 2.10

R&D / GDP 1.31 0.92 0.11 3.43

Market potential 28.86 0.64 27.40 30.13Note: The data appendix provides a detailed description of these variables.


ing on the mode of investment. In columns 4 to 6 in Table 4, the tax rate elasticities indicate a negative response of Greenfi eld investments (the reference group) to high taxes. Although the elasticities of the interaction term M & A × τ reported in columns 4 to 6 of Table 4 are positive and signifi cant (e.g., 0.304 in column 5), they cannot be necessarily interpreted as a clear indication that M&A investments react less sensitively to international differences in taxation than Greenfi eld investments. As stressed by Ai

Table 4Estimated Elasticities: Taxation and the Location Decision

(Full Sample)

(1) (2) (3) (4) (5) (6)τ –0.423* –0.477* –0.457* –0.777* –0.836* 0.793*

(0.125) (0.127) (0.129) (0.129) (0.131) (0.133)

M & A × τ 0.302* 0.304* 0.291*(0.020) (0.020) (0.021)

No. affi liates 0.067* 0.067* 0.057* 0.068* 0.068* 0.059*(0.008) (0.008) (0.008) (0.008) (0.008) (0.009)

Total assets 0.355* 0.355* 0.392* 0.360* 0.360* 0.396*(0.011) (0.011) (0.011) (0.011) (0.011) (0.011)

GDP capita 6.201* 6.412* 6.282* 6.266* 0.649* 0.634*(0.502) (0.508) (0.522) (0.504) (0.509) (0.523)

Population 11.77* 11.95* 11.91* 11.85* 12.05* 11.98*(0.506) (0.512) (0.522) (0.508) (0.517) (0.523)

Openness 0.265* 0.269* 0.277* 0.267* 0.272* 0.280*(0.039) (0.039) (0.040) (0.040) (0.039) (0.040)

Distance –2.958* –2.945* –2.747* –2.907* –2.891* –2.698*(0.172) (0.172) (0.174) (0.172) (0.172) (0.174)

Labor freedom 0.054 0.020 0.018 0.060 0.022 0.020(0.107) (0.107) (0.110) (0.107) (0.108) (0.111)

Corruption –0.289* –0.292* –0.282* –0.280* –0.284* –0.274*(0.084) (0.084) (0.086) (0.084) (0.085) (0.086)

Industry dummies No No Yes No No YesYear dummies No Yes Yes No Yes YesObservations 94,302 94,302 94,302 94,302 94,302 94,302No. of fi rms 2,627 2,627 2,627 2,627 2,627 2,627Log likelihood × 103 –10.27 –10.26 –9.935 –10.15 –10.14 –9.827

Notes: The table provides estimated elasticities. The underlying logit estimation results are not reported. Asterisks denote signifi cance at the 1% (*) level. Robust standard errors are reported between parentheses. All level variables are expressed in natural logarithms. The latent variable corresponding to the logit model yl = 1 is for the recipient country and yl = 0 otherwise. The statutory corporate income tax rate is τ. M & A is a dummy variable that takes the value 1 if the investment takes the form of M&A, and 0 if the investment is a Greenfi eld project. The data appendix provides a detailed description of the variables.


and Norton (2003), the interpretation of the estimated elasticity of the interaction term in non-linear models is not straightforward. Marginal effects in non-linear models are conditional on all independent variables included. To disentangle the estimated tax impact on M&A investments, the full interaction effect should be computed. The full interaction effect is a function of the cross-partial derivative of the expected value of the dependent variable. Hence, its statistical signifi cance depends on the signifi cance of the whole cross-derivative, and cannot be tested with the usual t-test on the interaction term.

To be sure, we compute the full interaction effect as a function of the predicted probability and the corresponding z-statistics. Figure 3 plots the results associated with specifi cation (5) in Table 4. The interaction effects are positive and signifi cant for almost all observations, confi rming the hypothesis that Greenfi eld investments react more strongly to high tax rates than M&A investments. The interaction effects are insignifi cant for few observations for the group of fi rms whose predicted probability is rather small (on the left side of the lower panel of Figure 3).

Additionally, we investigate whether or not the control variables may have different infl uences on the location probability depending on the mode of entry. Table 5 displays the resulting estimated elasticities. We concentrate the analysis on either the sample of Greenfi eld investments (columns 1 to 3) or the sample of M&A projects (columns 4 to 6). According to the results, for example, as indicated in column 1 of Table 5, we fi nd that an increase in the statutory corporate income tax rate of 10 percent (for instance, from 35 percent to 38.5 percent) reduces the probability of choosing a country to host a Greenfi eld investment by 6.41 percent. For the M&A sample, the elasticity of the tax rate is signifi cant but rather small. For instance, based on the estimates presented in column 4, an increase in the statutory corporate income tax rate of 10 percent reduces the probability of a country receiving a M&A investment by 2.78 percent, which is less than half of the tax elasticity of Greenfi eld investments (–6.41 in column 1). According to the meta analysis of de Mooij and Ederveen (2008), previous estimates imply a semi elasticity of –3.3; that is, a one percentage point increase in the tax rate reduces foreign investment fl ows by 3.3 percent on average across studies, but with a large variation. Assuming an average tax rate of around 30 percent, this translates into a tax-elasticity near unity. Our estimated elasticity for the full sample is around 0.45. As other stud-ies often use aggregate data, our below average elasticity may refl ect that the location decision is only one of several channels through which higher taxes lead to less FDI.

The results of Table 5 are consistent with the estimated positive interaction effects presented in Figure 3 and Table 4. Additionally, the negative M&A tax elasticity is consistent with recent studies that use aggregate fi gures on M&A (Di Giovanni, 2005; Coeurdacier, De Santis, and Aviat, 2009). However, our results suggest that M&A investments are less responsive to high taxes than are Greenfi eld investments.

With respect to the remaining control variables, the results reported in Tables 4 and 5 indicate that the presence of affi liated fi rms and their previous year total fi xed invest-ment in assets in a location increase the probability of entering this location again in all specifi cations. Additionally, the size (population) and the level of development of the


Figure 3The Interaction Eff ect as a Function of the Predicted Probability


host economy (gdp per capita) both have positive signifi cant effects on the odds ratio of the location probability. Distance has a negative effect as expected.9 The openness of the economy to international trade is a positive determinant of the location prob-ability of FDI in the whole sample. However, the sub-sample analysis indicates that the openness variable seems to play particularly a signifi cant role in the location decision

9 Hijzen, Görg, and Manchin (2008) examine in detail the role of trade costs for M&A activities.

Table 5Estimated Elasticities: Taxation and the Location Decision

(Sub-Samples)

Greenfi eld Sample M&A Sample(1) (2) (3) (4) (5) (6)

τ –0.641* –0.636* –0.557** –0.278*** –0.364** –0.324**(0.231) (0.234) (0.236) (0.153) (0.155) (0.157)

No. affi liates 0.036* 0.036* 0.040* 0.100* 0.100* 0.080*(0.012) (0.012) (0.013) (0.013) (0.013) (0.013)

Total assets 0.399* 0.399* 0.447* 0.323* 0.323* 0.364*(0.017) (0.017) (0.018) (0.015) (0.015) (0.015)

GDP capita 7.300* 7.419* 7.038* 5.867* 6.141* 6.097*(0.853) (0.863) (0.885) (0.643) (0.650) (0.669)

Population 14.23* 11.24* 14.17* 10.18* 10.48* 10.21*(0.966) (0.975) (0.984) (0.59) (0.600) (0.608)

Openness 0.541* 0.544* 0.575* 0.026 0.038 0.025(0.060) (0.060) (0.060) (0.059) (0.058) (0.060)

Distance –2.291* –2.306* –2.037* –3.362* –3.328* –3.132*(0.266) (0.266) (0.27) (0.232) (0.231) (0.234)

Labor freedom 0.110 0.133 0.084 0.021 –0.051 0.045(0.190) (0.189) (0.193) (0.132) (0.133) (0.138)

Corruption 0.066 0.103 0.074 –0.424* –0.457* –0.423*(0.134) (0.137) (0.141) (0.109) (0.108) (0.109)

Industry dummies No No Yes No No YesYear dummies No Yes Yes No Yes YesObservations 43,328 43,328 43,328 50,974 50,974 50,974No. of fi rms 1,306 1,306 1,306 2,321 2,321 2,321Log likelihood –3,998 –3,996 –3,883 –6,051 –6,039 –5,741

Notes: The table provides estimated elasticities. The underlying logit estimation results are not reported. Asterisks denote signifi cance at the 1% (*), 5% (**), and 10% (***) levels. Robust standard errors are reported in parentheses. All level variables are expressed in natural logarithms. The latent variable corresponding to the logit model is yl = 1 for the recipient country and yl = 0 otherwise. The statutory corporate income tax rate is τ. M & A is a dummy variable that takes the value 1 if the investment takes the form of M&A and 0 if the investment is a Greenfi eld project. The data appendix provides a detailed description of the variables.


of Greenfi eld investments. Although the labor freedom index has the expected positive sign in all specifi cations, it is insignifi cant.10 Further, as in Javorcik and Wei (2009), we fi nd that a decrease in the level of corruption, as captured by the Transparency Interna-tional index, increases the probability of selecting a location for an affi liate. However, this index loses its signifi cance in the Greenfi eld sample (columns 1 to 3 of Table 5), suggesting that different modes of investment may respond differently to corporate governance rules and the level of corruption in the host country.

IV. ROBUSTNESS ANALYSIS

In this section, we examine the robustness of our results. For example, high income economies may contain more attractive targets for acquisition than lower income economies. To address this issue, we split the sample into OECD and non-OECD coun-tries, and examine the effects of the corporate tax rate in both sub-samples. Columns 1 and 2 of Table 6 present the results. The estimated elasticities on the tax rate in both sub-samples are negative as expected, while the elasticities of the interaction term M & A × τ are positive. This fi nding confi rms our main results presented in the previ-ous section. The lower sensitivity of the location decision of M&A investments to differences in the tax rate is robust to the distinction between OECD and non-OECD countries. The estimated effect of corruption in the non-OECD sample is negative and its magnitude becomes larger in comparison to the results obtained from the full sample (Table 4). This indicates that a high level of corruption in a country, particularly in non-OECD host economies, reduces the likelihood that a new affi liate will be located in the country.

An additional question is whether the effect of the tax rate on location decisions is different for small investments than large investments. Based on the median of total fi xed and intangible assets of the new foreign investment, we distinguish between large and small affi liates. Columns 3 and 4 of Table 6 display the results obtained from both sub-samples, which do not suggest systematic differences between the responses of large and small affi liates.

As one could argue that a large plant expansion is similar to a Greenfi eld project, it is also of interest to account for potential effects of corporate taxation on plant expansion decisions. We exploit information on the balance sheets of German foreign affi liates in order to broaden the defi nition of Greenfi eld investments in our analysis. Specifi cally, in addition to new Greenfi eld entrants, we include a plant expansion, measured as an increase in the total balance sheet of an existing FDI affi liate of more than 50 percent, as a Greenfi eld investment. The results based upon this broader defi nition of Greenfi eld investments are reported in column 5. Expanding the defi nition of Greenfi eld investment does not alter the estimated tax effect.

In columns 6 to 9 of Table 6, we control for additional country characteristics. We include in column 6 the ratio of market capitalization of listed companies to GDP to capture the number of potential targets for acquisition. In a related issue, R&D consid-erations may trigger M&A activities. If high tax economies are characterized by high

10 Dewit, Görg, and Montagna (2009) fi nd that the level of employment protection deters inward FDI fl ows.

National Tax Journal832Ta

ble

6Es

timat

ed E

last

iciti

es: R

obus

tnes

s A

naly

sis

OEC

DN

on-

OEC

DLa

rge

Smal

lPl

ant

Expa

nsio

nFu

ll Sa

mpl

e(1

)(2

)(3

)(4

)(5

)(6

)(7

)(8

)(9

)(1

0)τ

–1.2

1*–0

.95*

–1.0

1*–1

.42*

–0.8

7*–0

.75*

–0.9

7*–0

.78*

–0.9

6*(0

.19)

(0.2

4)(0

.16)

(0.1

8)(0

.10)

(0.1

3)(0

.16)

(0.1

3)(0

.16)

M &

A ×

τ0.

30*

0.20

*0.

20*

0.26

*0.

13*

0.30

*0.

28*

0.29

*0.

28*

(0.0

2)(0

.04)

(0.0

3)(0

.02)

(0.0

1)(0

.02)

(0.0

2)(0

.02)

(0.0

2)

Effe

ctiv

e ta

x–1

.47*

(0.0

6)

Effe

ctiv

e 0.

26*

Ta

x ×

M &

A(0

.02)

No.

affi

liate

s0.

08*

0.06

*0.

03*

0.03

*0.

06*

0.06

*0.

06*

0.05

*0.

06*

0.07

*(0

.01)

(0.0

2)(0

.01)

(0.0

1)(0

.00)

(0.0

0)(0

.00)

(0.0

0)(0

.00)

(0.0

1)

Tota

l ass

ets

0.51

*0.

48*

0.46

*0.

41*

0.35

*0.

39*

0.39

*0.

39*

0.39

*0.

45*

(0.0

2)(0

.02)

(0.0

1)(0

.01)

(0.0

0)(0

.01)

(0.0

1)(0

.01)

(0.0

1)(0

.01)

GD

P ca

pita

11.1

*4.

76*

7.03

*6.

17*

6.80

*6.

24*

6.78

*6.

01*

6.28

*13

.1*

(1.2

0)(0

.95)

(0.6

8)(0

.76)

(0.4

2)(0

.53)

(0.5

6)(0

.57)

(0.6

1)(1

.17)

Popu

latio

n12

.9*

12.2

*13

.9*

10.3

*12

.6*

11.3

*12

.1*

11.5

*11

.5*

11.9

*(0

.70)

(1.0

3)(0

.71)

(0.7

2)(0

.42)

(0.5

5)(0

.58)

(0.5

7)(0

.68)

(0.7

9)

Ope

nnes

s0.

100.

37*

0.20

*0.

31*

0.30

*0.

18*

0.36

*0.

26*

0.28

*0.

32*

(0.1

0)(0

.08)

(0.0

6)(0

.05)

(0.0

3)(0

.04)

(0.0

4)(0

.04)

(0.0

5)(0

.08)

Dis

tanc

e–2

.74*

–3.3

4*–2

.94*

–2.6

3*–2

.64*

–2.7

4*–2

.39*

–2.7

6*–2

.53*

–0.8

0*(0

.23)

(0.3

8)(0

.24)

(0.2

2)(0

.14)

(0.1

8)(0

.19)

(0.1

8)(0

.20)

(0.2

9)


Labo

r 0.

03–1

.02

0.26

***

–0.1

60.

09–0

.04

–0.3

00.

00–0

.26

–0.4

4

Free

dom

(0.1

3)(0

.30)

(0.1

6)(0

.14)

(0.0

7)(0

.01)

(0.1

2)(0

.11)

(0.1

2)(0

.15)

Cor

rupt

ion

0.11

–0.5

6*0.

06–0

.57*

–0.1

0–0

.11

0.19

***

–0.2

8*0.

26*

–0.1

2*(0

.12)

(0.1

6)(0

.11)

(0.1

1)(0

.07)

(0.0

9)(0

.09)

(0.0

8)(0

.10)

(0.1

1)

Mar

ket

0.15

*0.

10**

C

apita

lizat

ion

(0.0

3)(0

.04)

R&D

/GD

P0.

39*

0.39

*(0

.04)

(0.0

4)

Mar

ket

2.23

*0.

77

Pote

ntia

l(1

.29)

(1.3

8)

Obs

erva

tions

32,3

7320

,957

51,1

8355

,948

141,

662

92,4

5081

,985

94,3

0281

,985

45,0

15N

o. o

f fi rm

s2,

532

1,09

51,

810

1,81

75,

137

3,61

63,

304

3,62

73,

304

2,83

1Lo

g lik

elih

ood

–5,7

32–2

,568

–5,1

62–5

,555

–14,

033

–9,7

65–8

,820

–9,8

25–8

,817

–6,7

80

Not

es: A

ster

isks

den

ote

sign

ifi ca

nce

at th

e 1%

(*),

5% (*

*), a

nd 1

0% (*

**) l

evel

s. R

obus

t sta

ndar

d er

rors

are

repo

rted

in p

aren

thes

es. T

he d

ata

appe

ndix

pr

ovid

es a

det

aile

d de

scrip

tion

of th

e va

riabl

es.


ratios of R&D expenditure to GDP, then the effect of the tax rate may to some extent capture R&D opportunities. We address such a possibility by adding the R&D inten-sity in the host economy to the set of controls. The results reported in column 7 show that the tax effect is robust to the inclusion of R&D intensity. In addition, one aspect that might drive the location decisions of multinational fi rms is market opportunities not only in the host economy but also in the region. We construct a proxy for regional market potential as the sum of the GDPs of the countries that are located in the same geographical region. We defi ne geographical regions based on the United Nations sta-tistical classifi cation. This market potential proxy is estimated to have a positive sign in column 8, but loses its signifi cance when we also include the market capitalization measure and R&D intensity (column 9).

While various studies fi nd that the statutory corporate income tax rate has signifi cant predictive power for the location decision (Büttner and Ruf, 2007), we also examine the impact of the effective tax rate on the location decision. Unfortunately, data on the effective tax rate are available only for 32 countries (mainly OECD countries). Accord-ing to our fi ndings in column 10, the impacts of taxation on the location decision of FDI fi rms are very similar for both tax rate measures.

Hence, the results of the various specifi cations presented in this section are reassuring, as a wide variety of robustness checks confi rm that the effect of corporate taxation on the location decision is signifi cantly higher for Greenfi eld investments than for M&A investments.

V. CONCLUSION

The empirical fi ndings reported in this paper contribute to our understanding of the role of taxation in determining the investment location decision. While most existing studies treat FDI as consisting of homogenous projects, our results distinguish between the effects of taxes on M&A and Greenfi eld investments. We have focused on the impact of corporate income taxation on the location decision of a German parent fi rm selecting a foreign host economy for a new entry. According to our fi ndings, an increase in the statutory corporate income tax rate of 10 percent reduces the probability of choosing a country to host a Greenfi eld investment by about 6.4 percent. M&A investments, however, are less sensitive to differences in international tax rates as indicated by a tax elasticity of –2.8 percent. This is consistent with (partial) capitalization of taxes in the acquisition price. Our results are robust to many sensitivity tests, including the use of the effective average corporate income tax rate instead of the statutory rate as the tax variable.

Since in the period of our sample all returns to FDI were exempt from German taxa-tion, we could not test the effects on investment decisions of the home country tax system. With regard to policy implications, our fi nding indicates that taxes affect the composition of FDI, as tax policies seem to affect differently multinationals’ location decision regarding M&A and Greenfi eld projects. This fi nding also contributes to a


growing literature on possible differential implications of the two modes of investment on the host economy.

ACKNOWLEDGMENTS

We are grateful to Heinz Herrmann, Beatrix Stejskal-Passler, Alexander Lipponer, and the research centre of Deutsche Bundesbank for their kind support. We thank Michael P. Devereux, Carsten Eckel, Clemens Fuest, Timothy J. Goodspeed, seminar participants at the Oxford Centre for Business Taxation, at the IIPF annual congress Uppsala 2010, and at the Max Planck Institute for Intellectual Property, Competition and Tax Law, Munich for their comments. We are grateful to two anonymous referees and the editors William Gentry and George Zodrow for helpful suggestions and comments. We are indebted to the ZEW Mannheim, especially Christoph Spengel and Christina Elschner, for providing data on the effective tax rates. The usual disclaimer applies.

REFERENCES

Ai, Chunrong, and Edward C. Norton, 2003. “Interaction Terms in Logit and Probit Models.” Economics Letters 80 (1), 123–129.

Andersson, Thomas, and Roger Svensson, 1994. “Entry Modes for Foreign Direct Investment Determined by the Composition of Firm-Specifi c Skills.” Scandinavian Journal of Economics 96 (4), 551–560.

Auerbach, Alan J., and Kevin Hassett, 1993. “Taxation and Foreign Direct Investment in the United States: A Reconsideration of the Evidence.” In Giovannini, Alberto, R. Glenn Hubbard, and Joel Slemrod (eds.), Studies in International Taxation, 119–148. University of Chicago Press, Chicago, IL.

Balsvik, Ragnhild, and Stefanie Haller, 2007. “Foreign Firms and Host-Country Productivity: Does the Mode of Entry Matter?” Working Paper. Norges Handelshøyskole, Bergen, Norway.

Becker, Johannes, and Clemens Fuest, 2008. “Tax Competition: Greenfi eld Investment versus Mergers and Acquisitions.” CESifo Working Paper No. 2247. CESifo, Munich, Germany.

Becker, Johannes, and Clemens Fuest, 2010. “Taxing Foreign Profi ts with International Mergers and Acquisitions.” International Economic Review 51 (1), 171–186.

Büttner, Thiess, and Martin Ruf, 2007. “Tax Incentives and the Location of FDI: Evidence from a Panel of German Multinationals.” International Tax and Public Finance 14 (2), 151–164.

Coeurdacier, Nicolas, Roberto A. De Santis, and Antonin Aviat, 2009. “Cross–Border Mergers and Acquisitions and European Integration.” Economic Policy 24 (57), 55–106.

De Mooij, Ruud, and Sjef Ederveen, 2003. “Taxation and Foreign Direct Investment: A Synthesis of Empirical Research.” International Tax and Public Finance 10 (6), 673–693.


De Mooij, Ruud, and Sjef Ederveen, 2008. “Corporate Tax Elasticities: A Reader’s Guide to Empirical Findings.” Oxford Review of Economic Policy 24 (4), 680–697.

Desai, Mihir A., and James R. Hines, Jr., 2003. “Evaluating International Tax Reform.” National Tax Journal 56 (3), 487–502.

Devereux, Michael P., 2007. “The Impact of Taxation on the Location of Capital, Firms and Profi t: A Survey of Empirical Evidence.” Oxford University Centre of Business Taxation Working Paper No. 07/02. Oxford University Centre of Business Taxation, Oxford, UK.

Devereux, Michael P., Christina Elschner, and Christoph Spengel, forthcoming. Effective Tax Rates in the EU, 1998–2007. Springer, Berlin, Germany.

Devereux, Michael P., and Rachel Griffi th, 1998. “Taxes and the Location of Production: Evidence from a Panel of U.S. Multinationals.” Journal of Public Economics 68 (3), 335–367.

Dewit, Gerda, Holger Görg, and Catia Montagna, 2009. “Should I Stay or Should I Go? A Note on Employment Protection, Domestic Anchorage and FDI.” Review of World Economics 145 (1), 93–110.

Di Giovanni, Julian, 2005. “What Drives Capital Flows? The Case of Cross-Border M&A Activity and Financial Deepening.” Journal of International Economics 65 (1), 127–149.

Djankov, Simeon, Tim Ganser, Caralee McLiesh, Rita Ramalho, and Andrei Schleifer, 2010. “The Effect of Corporate Taxes on Investment and Entrepreneurship.” American Economic Journal: Macroeconomics 2 (3), 31–46.

Egger, Peter, Simon Loretz, Michael Pfaffermayr, and Hannes Winner, 2009. “Corporate Taxation and Multinational Activity.” Oxford University Centre of Business Taxation Working Paper No. 09/04. Oxford University Centre of Business Taxation, Oxford, UK.

Head, Keith, and Thierry Mayer, 2004. “Market Potential and the Location of Japanese Investment in the European Union.” Review of Economics and Statistics 86 (4), 959–972.

Healy, Paul M., and Krishna G. Palepu, 2007. Business Analysis and Valuation: Using Financial Statements, 4th ed. Thomson Southwestern, Mason, Ohio.

Hijzen, Alexander, Holger Görg, and Miriam Manchin, 2008. “Cross-Border Mergers and Acqui-sitions and the Role of Trade Costs.” European Economic Review 52 (5), 849–866.

Hines, James R., Jr., 1999. “Lessons from Behavioral Responses to International Taxation.” National Tax Journal 52 (2), 305–322.

Huizinga, Harry P., and Johannes Voget, 2009. “International Taxation and the Direction and Volume of Cross-Border M&As.” Journal of Finance 64 (3), 1217–1249.

Huizinga, Harry P., Johannes Voget, and Wolf Wagner, 2008. “International Taxation and Takeover Premiums in Cross-Border M&As.” Oxford University Centre of Business Taxation Working Paper No. 08/26. Oxford University Centre of Business Taxation, Oxford, UK.


Javorcik, Beata S., and Shang-Jin Wei, 2009. “Corruption and Cross-Border Investment in Emerg-ing Markets: Firm-Level Evidence.” Journal of International Money and Finance 28 (4), 605–624.

Lipponer, Alexander, 2008. Micro Database Direct Investment –– MiDi A Brief Guide. Deutsche Bundesbank, Frankfurt am Main, Germany.

Markusen, James R., 2002. Multinational Firms and the Theory of International Trade. MIT Press, Cambridge, MA.

Miller, Terry, and Kim R. Holmes, 2009. 2009 Index of Economic Freedom. The Heritage Foun-dation and Dow Jones & Company, Inc, Washington DC and New York, NY.

Mintz, Jack M., and Alfons J. Weichenrieder, 2010. The Indirect Side of Direct Investment: Multinational Company Finance and Taxation. MIT Press, Cambridge, MA.

Modigliani, Franco, and Merton H. Miller, 1963. “Corporate Income Taxes and the Cost of Capital: A Correction.” American Economic Review 53 (3), 433–443.

Neary, J. Peter, 2007. “Cross-Border Mergers as Instruments of Comparative Advantage.” Review of Economic Studies 74 (4), 1229–1257.

Nocke, Volker, and Stephen Yeaple, 2007. “Cross-Border Mergers and Acquisitions versus Greenfi eld Foreign Direct Investment: The Role of Firm Heterogeneity.” Journal of International Economics 72 (2), 336–365.

Nocke, Volker, and Stephen Yeaple, 2008. “An Assignment Theory of Foreign Direct Investment.” Review of Economic Studies 75 (2), 529–557.

Raff, Horst, Michael J. Ryan, and Frank Stähler, 2009. “The Choice of Market Entry Mode: Greenfi eld Investment, M&A and Joint Venture.” International Review of Economics and Fi-nance 18 (1), 3–10.

Swenson, Deborah L., 2001. “Transaction Type and the Effect of Taxes on the Distribution of Foreign Direct Investment in the United States.” In Hines, James R., Jr. (ed.), International Taxa-tion and Multinational Activity, 89–112. University of Chicago Press, Chicago, IL.


Variable Defi nition Sourcey A dummy that takes the value 1 for the chosen host

economy and the value zero otherwiseDeutsche Bundesbank

τ The statutory corporate income tax rate Mintz and Weichenrieder (2010)

Effective tax The effective average corporate income tax rate Devereux, Elschner, and Spengel (forthcoming)

M & A A dummy that takes the value 1 if the investment takes the form of M&A and zero if the investment is a Greenfi eld project

Deutsche Bundesbank

No. affi liates The logarithm of the number of affi liates already operating in the host economy

Deutsche Bundesbank

Total assets The logarithm of total fi xed and intangible assets invested by the parent fi rm in location l

Deutsche Bundesbank

GDP capita The logarithm of gross domestic product per capita based on PPP of the host economy

IFS of the IMF

Population The logarithm of the total number of inhabitants in the host economy

IFS of the IMF

Openness The ratio of total trade (total imports plus total exports) of the host economy to gross domestic product of the host economy in current prices

IFS of the IMF

Distance The logarithm of the distance between Germany and the host economy

CEPII

Labor Freedom The logarithm of the labor freedom component of the Heritage index of economic freedom

Miller and Holmes (2009)

Corruption The Corruption Perception Index, redefi ned such that a high value of this index indicates a high level of corruption

Transparency International

Market capitalization

The ratio of market capitalization of listed compa-nies to GDP

WDI of the WB

R&D/GDP The ratio of government expenditure on research and development to GDP (percent)

UNESCO Statistics

Market potential

The logarithm of the sum of GDP of countries that are located in the same region. Geographical regions are defi ned based on the United Nations statistical database available from:http://unstats.un.org/unsd/methods/m49/m49regin.htm#ftna

IFS of the IMF and UN statistics

APPENDIX: VARIABLE DEFINITIONS AND SOURCES

Documents

THE EFFECTS OF TAXATION ON THE LOCATION · PDF fileNational Tax Journal, September 2011, 64 (3), 817–838 THE EFFECTS OF TAXATION ON THE LOCATION DECISION OF MULTINATIONAL FIRMS: